We have previously reported that flow cytometric tumor DNA content may be of prognostic significance in surgically resected small cell lung cancer (SCLC). We are particularly interested in determining prospective parameters for better selection of 'good prognosis' patients to proceed to surgery. Since flow cytometric measurements are poorly, if at all, applicable to endoscopic biopsy and cytology specimens we compared an image cytometric system to flow cytometry and clinical parameters in an extended series of surgically resected SCLC. Clinical follow-up was obtained on 75 patients having surgical resection for SCLC in the years 1981-92. Paraffin blocks were prepared for cytometry in standard fashion. Flow DNA histograms were characterised as diploid/tetraploid (n = 45) or DNA aneuploid (n = 27). DNA histograms obtained by image analysis were divided into type I (peridiploid n = 43) or type II (non-diploid with a minority of cells in peridiploid region; n = 31); 5c exceeding rate, 2c deviation index and malignancy grade were also computed. Overall two year survival was 27/75 patients (36%). Stage of disease was confirmed as a predictor of outcome with only 3/17 (18%) N2 patients surviving for 2 years as compared to 24/52 (46%) patients with N0/N1 disease. Image cytometric histogram classification just reached statistical significance with type I histograms indicating a better prognosis (20/43 survivors (47%) versus 7/31 (23%) patients with type II profiles, P < 0.05). Flow cytometry, 5cER, 2cD1 and malignancy grade were not useful in predicting prognosis. The results do not indicate a significant role for cytometry in SCLC.